computational social science: a literature review · computational social science: a literature...
TRANSCRIPT
Computational Social Science: A Literature Review
Anne Suphan & Christopher Zirnig
@czirnig
857 abstracts that quoted Lazer et al. (2009):
Introduction
New data are being obtained, generated and analysed for very different purposes. The
importance of big data, for example, for future social research is made clear in many writings.
For example, Conte (2016) refers to Big Data as the promise of large and relatively cheap
data sets as an opportunity for global data, focusing more on political decisions than on
potential profits. Especially unstructured “big data” will be a major challenge for social
science research methods. For example large amounts of texts can only be handled with
advanced computational methods. Creating these methods will require the knowledge and
theoretical background collected over the past 100 years in the qualitative social science
approach.
Method
After a systematic database search on “scopus" with the search word "computational social
science" we added 286 articles to our sample. From these 286 articles we drew a random
sample of 53 articles that are being categorised and analysed closely, using content analysis.
We will take a look at both established and quoted definitions of CSS and Big Data to try to
reach a level of abstraction, that will allow us to make statements about the similarities and
differences in the use of CSS and Big Data amongst the different scientific disciplines. We
also analyse the abstracts of all 286 articles with a cluster analysis. We then decided to
create a second dataset containing 857 articles that quoted Lazer et al. (2009) in order to
get a better picture in which fields the term “computational social science” is used.
First Results
We determined five broad categories: The first category deals with computer science,
neuroscience, biology and other sciences. It is about the development of artificial life and
especially artificial intelligence (Saunders / Bown 2015 and Díaz / Domínguez 2013). CSS is
used, for example, to enable machine learning (Hu et al. 2014 and Mason / Vaughan / Wallach
2014) and to better understand systems of multiple agents and man-machine
communications (Conte / Paolucci 2014 and Youyou / Kosinskib / Stillwell 2015). The second
category is about user-generated content from eg. social networks. This involves on the one
hand methodological questions such as datelining (Olmedilla / Martínez-Torres / Toral 2016)
and the development of new tools (Borra / Rieder 2014). Secondly, it also contains political
issues such as political online events in China (Liu / Zhong / Chen 2016). In the third category
we find forecasting and epidemiological inquiries such as Stohler (2014). The fourth category
deals with the development of social science theory in the course of CSS developments and
Big Data (eg. Shah / Cappella / Neuman 2015). The question is how social sciences are
changing and may need to reposition as (a) discipline(s) in line with technical innovations. The
fifth category is an "Others" category and summarises what is additionally discussed in the
discourse about CSS. For example gender as part of CSS (Purohit u. A. 2016), Policy-Making
(Lettieri 2016), new forms of public theories (Tufekci 2016) and urbanisation (Wei / Ye 2014).
Outlook
The definitions of CSS are broadly fanned and quite different and are still to be analysed and
categorised to find a common denominator. Cioffi-Revilla, for example, argues (2014):
"The new field of Computational Social Science can be defined as the interdisciplinary investigation of
the social universe on many scales, ranging from individual actors to the large largest groupings, through
the medium of computation." (P. 2)
Other studies define CSS as a better understanding of social behaviour (Strohmaier /
Wagner 2014). Still others put instrument based CSS in line with microbiology, astronomy or
nanoscience (Borra / Rieder 2014). Our main goal is to detect intersections among the field
of CSS and visualise opportunities for fruitful interdisciplinary approaches. For example Hu
et al. (2014) address the problem of dealing with large corpi of texts:
“Unlike information retrieval, where users know what they are looking for, sometimes users need to
understand the high-level themes of a corpus and explore documents of interest.” (P. 424)
This quotation makes it obvious, that the theoretical background of Levi Strauss, Ervin
Goffmann or Robert Park is highly needed in CSS methodology, dealing with unstructured
text data. Computational Social Science is an interdisciplinary field and therefore
cooperation is a necessity. We want to illustrate opportunities for cooperation. Topic
modelling and qualitative social research is only the beginning.
References:
Borra, Erik/Rieder, Bernhard (2014): Programmed method: developing a toolset for capturing and analyzing tweets., in: Aslib Journal of Information Management 66, S. 262–278.
Cioffi-Revilla, Claudio A. (2014): Introduction to computational social science: principles and applications, London ; New York: Springer.
Conte, Rosaria (2016): Big data: un’opportunità per le scienze sociali?, in: Sociologia e Ricerca Sociale, S. 18–27.
Conte, Rosaria/Paolucci, Mario (2014): On Agent-Based Modelling and Computational Social Science., in: Frontiers in Psychology 5, S. 1–16.
Díaz, Rodrigo Martínez/Domínguez, Álvaro Zapata (2013): Las ciencias sociales y los dispositivos de la complejidad., in: Cuadernos de Administración 29, S. 123–131.
Hu, Yuening/Boyd-Graber, Jordan/Satinoff, Brianna/Smith, Alison (2014): Interactive topic modeling., in: Machine Learning 95, S. 423–469.
Lazer et al. (2009) Computational Social Science. Science Vol. 323, Issue 5915, pp. 721-723.
Lettieri, Nicola (2016): Computational Social Science, the Evolution of Policy Design and Rule Making in Smart Societies., in: Future Internet 8, S. 1–17.
Liu, Tieying/Zhong, Yang/Chen, Kai (2016): Interdisciplinary study on popularity prediction of social classified hot online events in China, in: Telematics and Informatics.
Mason, Winter/Vaughan, Jennifer Wortman/Wallach, Hanna (2014): Computational social science and social computing., in: Machine Learning 95, S. 257–260.
Olmedilla, M./Martínez-Torres, M.R./Toral, S.L. (2016): Harvesting Big Data in social science: A methodological approach for collecting online user-generated content, in: Computer Standards & Interfaces 46, S. 79–87.
Purohit, Hemant/Banerjee, Tanvi/Hampton, Andrew/Shalin, Valerie L./Bhandutia, Nayanesh/Sheth, Amit (2016): Gender-based violence in 140 characters or fewer: A #BigData case study of Twitter., in: First Monday 21, S. 1–1.
Saunders, Rob/Bown, Oliver (2015): Computational Social Creativity., in: Artificial Life 21, S. 366–378.
Shah, Dhavan V./Cappella, Joseph N./Neuman, W. Russell (2015): Big data, digital media, and computational social science: Possibilities and perils., in: Annals of the American Academy of Political and Social Science 659, S. 6–13.
Stohler, Christian (2014): The Next Frontier: Digital Disease Detection in Cyberspace., in: Journal of Oral & Facial Pain & Headache 28, S. 105–105.
Strohmaier, Markus/Wagner, Claudia (2014): Computational Social Science for the World Wide Web., in: IEEE Intelligent Systems 29, S. 84–88.
Tufekci, Zeynep (2016): As the Pirates Become CEOs: The Closing of the Open Internet., in: Daedalus 145, S. 65–78.
Wei, Yehua Dennis/Ye, Xinyue (2014): Urbanization, land use, and sustainable development in China, in: Stochastic Environmental Research and Risk Assessment (SERRA) 28, S. 755–755.
Youyou, Wu/Kosinskib, Michal/Stillwell, David (2015): Computer-based personality judgments are more accurate than those made by humans, in: Proceedings of the National Academy of Sciences of the United States 112.
286 abstracts containing “Computational Social Science”: